Quick test of some cheap 50mm lenses

[Timber]

I've a collection of 50mm lenses and decided to run them against each other. Unfortunately I don't have access anything with very visible UV markings, the closest thing I can get was a pear :) as the brown and green parts of the pear reflects UV totally different, with the green parts being black, metallic, and the brown parts being matte greyish. I also added the UV absorbing flash screen from my Nissin Di622 and a box of Johnson's pure cotton buds as cotton is very reflective in UV. All was front of a white paper. So far these were the only things that I know can show UV characteristics, so please excuse me the "poor setup".

 

Settings:

All photographs were taken with the same setup (apart from the lenses) and settings Sony NEX6 with Spectrosil 2000 Fused Silica hotfilter replacement, f8.0, ISO 800, 1/60, Nissin Di622 Flashgun with 100% power, Hoya U360 + Schott BG40 filter stack.

To edit the pictures I used PhotoNinja with only color correction and cropping. The settings were copy-pasted, to ensure all photographs have exactly the same settings.

 

The Lenses:

Nikon EL-Nikkor 50mm f4.0 Old Version s/n: 284535 - 34.5 Filter Thread (FT), M39 - 44.5 Flange Focal Distance (FFD)

Nikon EL-Nikkor 50mm f2.8 Old Version s/n: 297262 - 40.5 FT, M39 - 43.5 FFD

Nikon Series E 50mm f1.8 Old Version s/n: 1186041 - 52 FT, Nikon F

Meyer-Optik Görlitz Domiplan 50mm f2.8 s/n: 3856193 - 49 FT, M42

E. Ludwig Meritar 50mm f2.9 s/n: 1551766 - 35.5 FT, M42

Jupiter-8 50mm f2.0 Silver version s/n: 6513399 - 40.5 FT, LTM/M39

 

The Pear:

Nicely shaped, picked from the local fruit store, sweet and tasty, a bit too hard. :D

The setup in Visible Light (made with an unmodified Sony A6000)

 

 

And now the results:

 

And a download link for all the RAW files ~90Mb:

http://clancode.hu/!uv/50mm.zip

 

My personal opinion: You can see the EL-Nikkor 50/f2.8 and the Meritar 50/f2.9 are making the biggest difference between the brown and green parts of the pear, with the Meritar giving a slightly brighter picture. The 50/1.8 Series E is very close as well and it has the best build quality. Personally I like the EL-Nikkor 50/f2.8 the most. I find it quite sharp and easy to use, but since the short FFD you can only have focus to infinity if you're using one of the Mirrorless Cameras wtih short FFD. The 50/f2.0 Jupiter-8 performs the worst, and it has all of its coatings removed so I can conclude it is not really suitable for UV photography.

 

All lenses are easy to find on eBay and go for low price, so any of these lenses (apart from the Jupiter-8) is good for beginners. For Nikon I would recommend the 50/1.8 Series-E as it's Nikon's F mount so no adapters necessary. For any other DSLRs I'd recommend the Meritar 50/f2.9 for the slightly better transmission and for Mirrorless cameras you can either go for the Meritar or the EL-Nikkor 50/f2.8 as the later seems to me a little sharper.

[Andrea B.]

Timber, thanks for the nice test of the 50s. We all love to see how the various UV-capable lenses perform. And it is a nice touch to include a zip of your raws.

 

The flash screen and cotton buds were a good idea! Sometimes cotton or paper or other white materials can fluoresce under UV because they contain optical brightners. But I suspect that cotton buds do not contain such brighteners given their use.

 

Next time, if you want a "standard" UV test subject, you might like to go to a store where flowers are sold and look for a Sunflower. All Sunflowers show some kind of dark UV-bullseye center which extends onto the petal bases by different amounts depending on the particular Sunflower species. I don't know where flowers are sold in London. Here in the US Sunflowers can be found at either florist shops or in the floral section of grocery markets or stores.

 

[Timber, you might have noticed that we do sometimes make very minor edits for clarity or spelling so that UVP posts can be referenced for future use.]

[Andrea B.]

The Jupiter 50/2 does seem to produce the noisiest photo - likely because it needed another 1/2-1 stop of exposure. From the EXIF I see that the Jupiter was shot at 1/80" and the others were shot at 1/60". That is only 1/3 stop difference and probably does not fully account for all the extra noise in the Jupiter shot. Probably the Jupiter simply transmits less UV.

 

The EL-Nikkors 50/2.8 and 50/4.0 seem sharpest. Not surprising at that distance given that they are enlarging lenses.

 

The 50/1.8 Nikon E seems a bit soft. But I don't want to judge that from just one photo. "-)

 

The Domiplan 50/1.8 and Meritar 50/2.9 seem about the same. Again, seeing just one image from each makes it difficult to decide between them. (And I was not sure where your focus point was. In one shot the background is sharp and print less sharp. Vice versa in the other.)

 

The white balance is all over the map. I was curious whether a more uniform wb could be achieved. Of course the EL 50/2.8 has a mind of its own about colour as already shown in one of your previous posts. "-)

 

I note in passing that the NEX-6 has a lot of headroom to pull back the blown highlights successfully as you can see in the cotton tips. The shadows also open up a bit, although you cannot see that after the resizing. Good to know this about NEX highlights/shadows as I'm interested in a NEX (or alpha as they are called now).

 

[Timber]

Hi Andrea,

 

I don't mind if my posts are being edited for proper grammar... my mother tongue is Hungarian and not English so I am sure I make loads of mistakes. :)

 

Here at Finsbury Park I only know about 10 fruit store but can't think of any flower shops :) And at the £ store they only have plastic flowers :D The only reason why I choose a pear because it shows some patterns in UV, and hey, it's something new, something fresh (literally :D)! My goal with test was to see which cheap lens could be the best for someone to start with, as I believe noone can really afford to have the Coastal Optics as their first UV lens and based on my experience it's really frustrating when your lenses are limiting you... and since I already went through the acquire-test-disappoint phase with these lenses I thought to share it so others don't have to :) They can just go to eBay and grab one of the cheap lenses and start their journey into the UV photography :) On the whitebalance and editing... I wanted to see that with exactly the same settings how they compare against each other that's why the WB was so crazy (actually it was the same on all... which gives you the different results). Also I purposely did not edit anything on the photos so it will show what the lens can really do without post processing (I think with the Jupiter I accidentally changed the shutter speed... oops :P).

[Andrea B.]

I liked the Pear. It might be interesting to cut one in half and see how that looks in UV.

I encourage everyone to make portraits of everything so that we can see how it all looks in UV!!

 

It is wonderful that there are some inexpensive UV-capable lenses available. We have quite a nice list to which everyone has contributed over time. I'll be updating it later this fall when it begins to get cold and I want to work indoors.

[enricosavazzi]

A general comment is that false-color yellow in UV images is usually produced by radiation around 350-360 nm. The fact that you have more yellow in the Meritar sample (I assume you used the same WB in all pictures, and are not using auto WB) suggests that this lens is transmitting deeper into the UV than the others. These wavelengths are transmitted by both U360 and BG40, and the U360 does not transmit VIS yellow, so we should be safe from VIS contamination. NIR contamination should affect all lens samples, so it is not likely to be an explanation for the greater yellow in the Meritar sample. We can also exclude all wavelengths shorter than 310 nm, because neither U360 nor BG40 transmit enough of them.

 

The bluish tone of the Jupiter 8 is consistent with UV radiation very close to the VIS range, perhaps 395 nm, or even some contamination from VIS blue and violet. Depending on WB, UV at these wavelengths registers as blue or violet, or even mostly red if WB is very off. This suggests that this lens transmits less far into the UV than the others. This is also consistent with the fact that it seems to require a longer exposure.

 

So my bet for selecting a UV lens among these would be on the Meritar, which seems better than the El-Nikkors. Spectral transmission is only one of the main criteria, and you need to test for UV "focus shift", which can be a problem with broadband UV-pass filters, and general image sharpness and contrast in the UV.

 

Edit - If you started using this electronic flash only recently, it may be a good idea to shoot a series of multiple images with the same lens and setup, just to make sure there are no random variations in spectral emission from shot to shot. One possible source of variation is the time allowed for the flash to recycle. If the capacitor is not fully charged, the emission will not only be shorter/weaker, it may also be spectrally skewed toward shorter wavelengths.

[Andrea B.]

That's an important reminder about the possibility of variation in spectral emission from a UV flash.

 

I do note, however, that the false colours we obtain are in some small way dependent (among several dependencies) on the raw converter used. This is one reason why it is always difficult for me to say that a certain false colour can be attributed to a specific wavelength when I see that colour from one converter and not another. (I'm thinking of that dark green I see sometimes, for example.) In general the principle seems to hold, but it could be misleading sometimes.

[enricosavazzi]

[...]

I do note, however, that the false colours we obtain are in some small way dependent (among several dependencies) on the raw converter used. This is one reason why it is always difficult for me to say that a certain false colour can be attributed to a specific wavelength when I see that colour from one converter and not another. (I'm thinking of that dark green I see sometimes, for example.) In general the principle seems to hold, but it could be misleading sometimes.

That is absolutely true, comparing among different raw converters (and even different installations and settings of the converters, and different cameras and their WB setups) is always problematic.

 

In this case, we have a sample of images shot with the same equipment and camera settings, and presumably post-processed in the same way, so we can risk making a (relative) comparison among the lens samples and be reasonably sure that the differences we see are mainly due to the lens (save for possible quirks like fluctuations of flash emission). In principle, a more yellow image than other samples made with different lenses (and all other conditions remaining the same) likely indicates a higher relative amount of this color - albeit not infallibly so, as you point out.

[Andrea B.]

Enrico, I'm probably overthinking it. :D

 

I've just always been a bit puzzled by that dark green. Sometimes I sure do wish I had some measuring equipment.

[Timber]

I was a bit curious on how a painting will look like in UV/VIS/IR and used the Meritar and the EL-Nikkor f2.8. My finding is that the Meritar has a quite big focus shift (will do proper tests later) and the EL-Nikkor just a minor one. Also did the Brick Wall Challenge with the lenses :D and the EL-Nikkor is running circles around the Meritar in terms of sharpness in the Visible range (have not tested in UV / IR... yet! :D)

 

Also I have some black-light bulbs at home and will try to get some better UV models than the pear and do the test again (with the right safety measures - suncream and googles). And I also plan to do a test with coca cola as I know it's entirely transparent in IR, but wonder how it will look like in UV... or Tonic! :P Oh, the curiosity of a child I have :)

[Andrea B.]

We have one painting here:

http://www.ultravioletphotography.com/content/index.php?/topic/735-oil-painting-in-uv-vis-ir/

 

Hope to see more of this.

[enricosavazzi]

Andrea, can you point me to an example of the false-color dark green you mean?

[JCDowdy]

Enrico,

 

I recall this recent discussion on the topic of UV-Green.

 

I sometimes see a yellow or green tint which I suspect is a WB issue but I have not tried to intentionally reproduce the effect.

I do agree with you that images skewed to the yellow or blue can be the result of WB bias to shorter or longer wavelengths relative to the WB.

This was recently clearly demonstrated by a comparison of different UV-LED flashlights (or torches as they are otherwise known). Those of us who have stored separate custom WB for sunlight and UV-LED and subsequently accidentally selected the wrong WB have experienced a similar effect. :D

 

My feeling is this yellow/green false color anomaly seems to occur more easily with some lenses rather than others. Would a lens with a transmittance cut off in the yellow/green region of the monochromatic false UV-color spectrum possibly be be more prone to this false color anomaly?

[enricosavazzi]

I cannot directly answer your question because we don't seem to have all data/facts yet. As long as we use broadband UV-pass filters, there is always going to be uncertainty as to which wavelengths record as which false color on Bayer sensors.

 

I have done some work with narrowband (10 nm) filters, but only in the lower UVA wavelengths. These filters are expensive, so I may eventually get some for higher and even lower wavelengths, but this does not have a high priority. Nonetheless, what I can say (with different degrees of certainty, as noted below) is:

 

350 and 360 nm record in both the green and red channels (high certainty). Depending on WB and intensity, this is rendered as yellow or orange, or sometimes even red.

 

340 nm records in the green channel with a slight amount recorded in the red channel (high certainty). WB plays only a minor role in the appearance of this false color, unless the red channel is substantially pumped up.

 

325 nm records practically only in the green channel (high certainty). Because of the purity of this color, it is essentially unaffected by WB.

 

300 to 305 nm seems to record in the blue channel (low certainty). I have however failed to see an experimental confirmation of this in test done with a monochromator at these wavelengths, suggesting that this is not true. The blue in this case may have other causes, like demosaicking, surface iridescence, or non-linear sensor response. The tantalizing fact is however that it is visible in only a minority of flower species. A sensor or camera artifact should always be present regardless of subject. Klaus has published images that show this blue with some subjects and with filters than pass also UVB, which he interpreted as high UVB reflectance by the subject. I get the same blue with UVB-pass filters and some flowers, but it does behave strangely (in most cases, more like a mirror-like reflectivity or iridescence than "solid" color, although in very few cases I do get a "solid" blue not highly dependent on illumination and viewing angles).

[Andrea B.]

Can our cameras record UVB? I did not know this. I guess that surprises me.

 

**********

 

The dark green I'm referring to is not from foliage. Green is quite common in UV landscapes in some leaves or grass. For foliage, I've gotten dark-grey, dark-yellow-green, black, dark grey-green and dark-blue-grey foliage in various photographs.

 

The dark green which puzzles me seen instead in some UV dark flowers.

 

I've also noticed that the false colours I get in closeups do not always appear quite the same in landscapes where the photograph is made at a much longer distance from the subject. It is the UV "fade" I suppose. But yet another reason why we must be careful about what we say the false colours mean spectrally.

 

There is a LOT of imprecision in raw conversion. One little twist of the converter and you get false colours just different enough to make one doubt that there is any certainty in this little game at all.

 

Any statements about the spectral meaning of false colour needs to be tested across many different cameras and lenses. We do not really have that much data at hand so far. Ideally, by gathering UV floral signatures here on UVP we will gather some useful data towards various ends including, possibly, false colour interpretation. No one is going to convince me yet of anything by exhibiting photos only from one converted camera!! But I am convincible if the supporting data are there.

 

As a side note, I have made thousands of flower photographs to date, and I almost never get the camera to properly record the violets and purples. Even with CCPassport and camera profiles and setting in-camera white balance. So I tend to have my doubts about how precisely any converted camera is recording the UV "colours". :D

 

I will look for one of the flowers I'm talking about and post it here.

[Andrea B.]

It is probably time for me to pony up for a set of good UV narrowband filters. :D

Maybe I can get a loan from the National Science Foundation. .

I was going to wait on that until I got the NEX. Then I would have 3 cameras to work with.

 

However in the world did I get involved in such an expensive "hobby"???? :P :) :) :D

[enricosavazzi]

Can our cameras record UVB? I did not know this. I guess that surprises me.

[...]

I summarized my findings at http://savazzi.net/p...raphy/uv-b.html

 

However, at present I am not as sure about the conclusions as I was when I wrote this page. An earlier post in this thread reflects my current thoughts better. Whatever it is that my camera is recording, it is at very low levels.

 

Update: I have a special filter on order that may help to shed light (pun intended) on this question (hint: not a narrowband filter, quite the opposite actually). It will take a few weeks for me to come back to this question. When I do, I should have a definitive answer to whether this "phantom blue" or "dead blue" is UVB or something else. In the latter case, exactly what else, however, will have to be investigated.

[Andrea B.]

This sounds very intriguing, Enrico. We will be eager to read about your experiment.